What The Solyndra "Scandal" Says About Government and Innovation

It's pretty clear that the Solyndra bankruptcy has been blown out of proportion by Republican politicians looking for a great attack line. But this extremely minor event is interesting because it helps us understand the conflicting incentives faced by many government programs. As many commentators have pointed out, the bankruptcy is no big deal: firms like Solyndra receive federal assistance precisely because their footing in the market is (at least initially) shaky. In fact, a company that exists at the limit of technological innovation going out of business could even be seen as a positive indicator; a sign that government is taking risks.

Critics are not without a point, however. As a general principle, wasting millions of dollars is never a good idea, and sinking $535 million of loan guarantees into a company that goes bankrupt provides a pretty decent prima facie case for asserting a government boondoggle. But consider the implications of jumping immediately to criticism. Are we really prepared to say that government should never, ever, spend money on stuff that might fail miserably, even when there's a chance of a massive positive outcome?

All organizations must manage the tension between cost-effectiveness (which encourages conservatism and status-quo practices) and innovation (which increases productivity and encourages experimentation). This tension is especially vexing for government because criticism is so intense in both directions. If government doesn't innovate (or use cutting-edge technology), it's a calcified mass of bureaucratic inefficiency. But when new programs or institutions fail (as they necessarily will), spendthrift government squandered our money on ridiculous pipe-dreams. Generally the second outcome is perceived as worse (it certainly makes a bigger media splash), which breeds an unhealthy level of risk aversion among politicians and civil servants.

So what can we do? Well, unfortunately there's no real solution on the level of a single organization, but there are ways to get around the problem using a more systems-minded approach. The free market innovates because one big success can more than make up for several past failures. But this solution only works because failures disappear pretty quickly, which frees up capital for new experimentation.

Government's real problem is not that it loses some money on bets that don't pay off (like Solyndra); the potential rewards are well worth the risk and on the whole public support for innovation is a good idea. The problem is with businesses like Solyndra who receive government assistance and don't go bankrupt. Zombie assistance programs that continue to draw resources with no probability of big future rewards are the real culprit. Unfortunately it takes time to identify such programs, by which time vested interests have entrenched themselves around the new status quo.

Still, this gets to the core of the solution: government programs should be more willing to subsidize research and innovation, but--critically--be more willing to cancel funding once a program's long-term survival ability without government support becomes known.

There are exceptions, of course. If the magnitude of a potential reward is large enough, more loyal assistance programs might be called for. Nuclear fusion energy research is a great example: the joke that fusion is 30 years away and always will be is true--until it's not. The point is that for innovation, government should adhere to the principle of cost-benefit analysis, but only in the aggregate and only in the longer run. So don't complain about government boondogglery when one subsidy program proves to be a waste of money. It's the system that counts.